What does "overfitting" mean in machine learning?
1) The model is too complex.
2) The training process is too fast.
3) The model performs well on the training data but poorly on the new data.